ESDS Launches Sovereign GPU Service to Power India’s AI Ambitions
Bengaluru: ESDS Software Solution Limited has launched a sovereign-grade GPU as a Service (GPUaaS) platform, a major strategic expansion announced at its 20th Annual Day in Nashik. The service, built on next-generation GPU SuperPODs, is designed to meet the surging demand for high-performance computing to fuel AI, Generative AI, and Large Language Model workloads across enterprises, BFSI, research, and government sectors.
This move positions ESDS as an end-to-end provider of cloud and AI infrastructure at a time when global spending on AI-optimized servers is projected to hit $329.5 billion by 2026. The managed service offers purpose-built, secure, and low-latency environments for mission-critical AI tasks, eliminating the complexity and high cost that often hinder organizations.
Piyush Somani, Promoter, Managing Director and Chairman of ESDS, stated, “This is a strategic leap forward to address the surging demand for large-scale AI infrastructure. For far too long, organizations have been held back by the complexity and prohibitive cost of GPU infrastructure. With this launch, we are democratizing access.”
The service leverages cutting-edge hardware, including NVIDIA’s DGX and HGX B200, B300, GB200, and the NVL72 architecture, as well as AMD’s MI300X platforms. These SuperPODs are engineered with high-bandwidth NVLink and AI-optimized orchestration for predictable performance. ESDS’s full-spectrum portfolio includes everything from advisory and deployment to a fully managed cloud, handling rack engineering, networking, and 24/7 MLOps.
A key feature is the SuperPOD Configurator, a tool that lets enterprises design their AI infrastructure, generating instant performance estimates and transparent cost projections. The service’s impact is demonstrated by a research lab that cut its 50-billion-parameter model training time from over 40 days to just 10, while reducing costs by 60% and achieving 30x faster inference by transitioning to ESDS’s NVL72-based infrastructure.
